Introduction
In the 1970s believed that people are generally rational, with strong emotions being the cause departure from rationality. Since then behavioral psychologists and economists have developed a different view based on an analysis of how we think. TFS describes two systems of thought that synthesizes this understanding of the mind’s dual operations: System 1 and System 2. These two Systems play the role of distinct characters in the book, with unique personality attributes who both cooperate yet compete for control of our attention and behavior.
To isolate the behavior of System 1, consider the picture
images.jpeg
In fairly effortless manner, we are able to surmise several facts about this woman: she is angry and quite likely to say unkind words in a loud and harsh voice. This assessment came to mind instantaneously without effort, and is an example of fast thinking.
On the other hand, consider the question
17 X 24 =
We can immediately identify this as a multiplication problem, and are aware there exists a paper and pencil solution. We can recognize that 12,000 is too big and 123 is too small to be the answer. However, without further thought it is unlikely that we can surmise the correct solution was not 568. In order to carry out the analysis a different mode of thinking needs to ensue – slow thinking. You will retrieve from memory a cognitive program learned in school for carrying out multiplication. You effortfully need to carry out a sequence of steps, retaining numbers in working memory to hold onto intermediate results which feels like a temporary strain. The overall process of mental work is deliberate, effortful, and orderly, a prototype of slow thinking. The effects of slow thinking were not only felt in the mind – it had physical impact on you – your muscles would have tensed, blood pressure rose, heart rate increased, and pupils dilated, which all would have relaxed upon when you found the right answer.
System 1 and System 2
Other examples of System 1 and System 2 in action include:
Table 1
| System 2 | System 1 |
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
As TFS describes it
“System 1 operates automatically and quickly, with little or no effort or sense of voluntary control”
*“System* 2 allocates attention to the effortful mental activities that demand it, including complex computations. The operations of System 2 are often associated with the subjective experience of agency, choice, and concentration”.
The key theme of the book is that whereas our self-identities (and the ideal aspects of Homo Economicus) are bound in the version of our System 2 selves, System 1 in fact drives the lion’s share of our everyday thoughts and actions. A key focus of the book then is to investigate the specific tendencies of System 1 to reach judgments and make choices which are fast, often intelligent, but exhibit systematic biases. The other question the book considers are the capabilities of System 2 that are directed towards control and the correction of the automatic processes of System 1.
The key distinguishing characteristics between System 1 and System 2 can be describe more fully as follows.
| System 1 (Intuitive) | System 2 (Reflective) |
| Process Characteristics Automatic Effortless Associative Rapid, parallel Skilled action |
Process Characteristics Controlled Effortful Deductive (formal rules of reasoning) Slow, serial Rule application |
| Content Characteristics Affective (emotional) Causal propensities (relationships) Concrete, specific Prototypes |
Content Characteristics Neutral Statistics Abstract Sets |
Heuristic Reasoning Homo Economicus
The Stroop Task
A useful way to recognize the simultaneous joint existence of these two systems of thinking, consider the “Stroop Task”:
Direction: State the color in which each word is printed.
Set 1: elf, ball, table, storm, group, end, find, away, drop, book, > fish,
> tree, ...Set 2: red, blue, red, red, red, blue, blue, blue, red, blue, blue, > red, red,
> red, ...Set 3: red, blue, red, red, red, blue, blue, blue, red, blue, blue, > red, red,
> red, ...
The “Stroop effect” is that Set 3 is much slower than Set 1 or Set 2 to complete. Why is Set 3 harder than Set 1 or Set 2? We can understand the effect through the joint operations of System 1 and System 2.
An explanation of the Stroop Effect based on the above schematic is as follows. Reading is a highly practiced, automatic skill. System 1 thus accesses the word meaning (concept of red
or concept of blue) quickly and automatically. This process is involuntary and cannot easily be suppressed or inhibited. The task instruction itself (in the directions) requires following a rule: “Name the font color”. The System 2 rule processing is slower than the automatic reading process of System 1. There is a response conflict between saying "blue" when looking at a blue "red". Also, there conflict between saying "red" when looking at a red "blue". The response conflict in turn causes slower response.
Stroop effect exemplifies two concurrent processes of thinking:
Automatic process extracting the meaning of “blue”
Controlled process identifying the color of “blue”
Reading the word meaning cannot be voluntarily inhibited and a difficulty occurs when an automatic process (extracting the meaning of a color word) conflicts with a rule-governed process (naming the color). A key idea is that the automatic and controlled processes can both occur at the same time, and sometimes they come into conflict. Herein will lie a key source of tension that cause human departures from Homo Economicus: Heuristic reasoning conflicts with rule-based reasoning and can sometimes dominates.
Gorilla Experiment
Another phenomenon that reveals is the tension between System 1 and System 2 is this famous experiment by Chabris and Simmons. The experiment shows that both System 1 and System 2 compete for the finite attentional resources of the mind. These attentional resources can be associated with the “working memory” module in the schematic model of cognitive resources below (taken from Hastie and Dawes, 2010). For System 2 to perform, it has to assume resources of working memory (through its executive function and control), threby limiting the capacities available for System 1.
Interaction between System 1 and System 2.
How are System 1 and System 2 linked? TFS explains the connection rather astutely (p. 24):
In the story I will tell, System 1 and 2 are both active whenever we are awake,. System 1 runs automatically and System 2 is normally in a comfortable low effort mode, in which only a fraction of its capacity is engaged. System 1 continuously generates suggestions for System 2: impressions, intuitions, intentions, and feelings. If endorsed by System 2, impressions and intuitions turn into beliefs, and implulses turn into voluntary actions. When all goes smoothly, which is most of the time, System 2 adopts the suggestions of System 1 with little or no modification…. When System 1 runs into difficulty, it calls on System 2 to support more detailed and specific processing that may solve the problem of the moment.
Examples of problems that activate System 2’s involvement include all those listed on Table 1, which can range from solving a math problem, exhibiting self-control in specific social situations, heightened awareness/concentration when driving at night, or when an anomaly/risk presents itself suggesting an error is about to be made (among others). The common pattern is that some friction in System 1’s usual function is an invitation of System 2 to act.
There is one massive problem – System 2 is lazy! Given the choice, it errs on the side generally (although with considerable individual heterogeneity) towards inaction, e.g., it abides by the “law of least effort”. To appreciate this tendency towards a lazy state, refer back to the cognitive schematic – the control of the attention and working memory by System 2 for deliberative thinking requires effort and work. Thus both self-control and cognitive effort are forms of mental work. This leads to a phenomenon of ego-depletion: “activities that impose high demands on System 2 require self-control, and the exertion of self-control is depleting and unpleasant.” Difficult System 2 work, whether cognitive reasoning or self-control, in-fact consumes blood glucose.
In a remarkable study published in the Proceedings of the National Academy of Sciences on Israeli judges shows some perverse consequences of ego-depletion. In this legal setting, there are parole judges reviewing applications for parole, and each application is considered for 6 minutes each. The default decision is a denial (so they are deliberating to overturn a denial). On average 35% approved, however 65% approved immediately after a meal which dwindles to nearly 0% before next meal. The pattern is depicted in the findings from the paper below.
There is considerable individual heterogeneity in the natural tendency of System 2 to exert effort for executive control over System 1. Consider one of the more famous experiments in the history of psychology: the Walter Mischel cookie/marshmallow experiment.
This suggests there are at least two different dimensions of System 2 discussed by Stanovich and West in their book Rationality and the Reflective Mind. One is intelligence – the sheer ability to excel in slow thinking and computation. The other is rationality, which is challenged by the degree to which System 2 wants to be lazy, which also varies at the individual level and features of what the authors call the reflective mind.
The effort expended by System 2 in relation to the effortless automation of System 1 provides some understanding as to why human cognition evolved this way? As TFS explains (p. 25)
The division of labor between System 1 and System 2 is highly efficient: it minimizes effort and optimizes performance. The arrangement works well most of the time because System 1 is generally very good at what it does: Its models of familiar situations are accurate, its short-term predictions are usually accurate as well, and its initial reactions to challenges are swift and generally appropriate. System 1 has biases, however, systematic errors that it is prone to make in specified circumstances. One further limitation of System 1 is that it cannot be turned off.
The interaction can be described graphically as followed
System 1 is an “Associative Machine”
A major feature of System 1 is the ability/speed to function as an “associative machine”, e.g., form links between ideas/actions/events in associative memory and develop causal and coherent accounts of these linkages. Thus, for example if you see a dog, you expect to see his master. Alternatively, if you hear a breaking sound, you expect to see that something has been broken. Certain effortless instincts and associations of System 1 are learned automatically from birth, e.g., detect hostility in a voice.
Other automatic activities of System 1 become learned through prolonged practice that creates associations between ideas, e.g., the expertise of a chess master in reading a board. From this network of associations stored in memory, System 1 can create inference/understanding/accounts of sensory data in the environment in automatic and instantaneous fashion. As TFS describes it “the knowledge is stored in memory and accessed without intention and without effort”.
As an example of the associative phenomena, consider the following example of two words on a page:
Banana Vomit
One of the first things System 1 experience upon reception of these words is a mild element of surprise – it is uncommon to see these words juxtaposed in this fashion (perhaps having never seen together at this point). Nevertheless, the associative machine will immediately work to tie these two together using properties of associative memory. The most likely association it will create is a temporal sequence whereby the banana caused the sickness inducing the vomit. This gravitation towards associative coherence will take place automatically without our conscious intervention or effort.
Moreover, once this association is now formed and placed (at least temporarily) in memory, it will automatically activate a cascading network of related associations that can affect our behavior- a phenomena known as priming. In this case, we will feel a momentary physical aversion to bananas, and we will be more attuned receptive to concepts related to “vomit” and “bananas” – sick, stink, nausea, yellow, fruit, apple, berries.
The mechanisms of association of ideas should be seen as nodes in a large network (associative memory) in which each idea is linked to many others. The linkages can take have different underlying meaning/interpretation such as
Causes linked to effects (virus cold)
Things linked to their properties (lime green)
Things to the category in which they belong (banana fruit)
The association of ideas was understood by Enlightenment philosophers John Locke and David Hume as a basis for human intelligence. The modern scientific appreciation to the subject adds to this philosophical tradition by recognizing that each when a node in this network is activated in memory, it triggers activation of the entire linked network of ideas/concepts all at once in a manner that is automatic (System 1) and hidden from our conscious selves. The activation mechanism is symmetric - causes activate effects and effects activate causes.
Priming
The functioning of this associative machinery leads to a certain “manipulability” of System 1 through associative cues known as priming. Priming arises when exposure to one idea makes an individual more receptive to the suggestion of another idea due to familiarity induced by change in the state of memory from the network of underlying associations. A mundane illustration of priming is when exposed to “eat” it makes it more likely someone fills in the blank of SO_P with SOUP rather than SOAP (and vice versa if exposed to “wash” rather than “eat”). However, there are some more non-trivial manifestations of priming that Chapter 4 documents.
Among the most famous of these is the “Florida effect”. John Bargh at NYU conducted a study with students (aged 18 – 22) who were asked to assemble 4-word sentences from 5 words Eg: “finds he it yellow instantly”. The treatment group was asked to unscramble sentences that included words such as
Florida, forgetful, bald, grey, wrinkle
The students were then asked to walk down the hall to complete another task, but it turned ou this walking trip was the heart of the experiment. The key finding is that the treatment group walked down significantly slower than others.
The “Florida effect” involves two stages of priming. First the words exposed to the treatment effects primes ideas of old age (observe the word “old” is never mentioned). Second, old age primes a behavior, namely walking slowly. An important dimension of the hypothesized effect is that the priming mechanism also works through the reciprocal links of the associative machine. A study conducted at a German university showed the Florida effect in reverse. Students were asked to walk around a room for 5 minutes at 1/3 of the normal pace (30 steps per second). Afterwards they were much quicker to recognize words associated with old age. Likewise, in another study, tests were done while subjects are smiling/frowning (induced by a pencil held in the mouth) or nodding/shaking head. Depending on the treatment, these groups differed in finding cartoons funnier, upsetting pics worse, and radio editorials more or less agreeable.
Priming effects play a major role in how firms attempt to market products to consumers. More recently, priming effects have launched theories and practices as to how to induce specific influences and behaviors in others through “pre-suasive” cues.
Endorsing Intuitions: Cognitive Ease
When System 1 can readily achieve cognitive coherence, it provides a sensation of cognitive ease. It is lack of cognitive ease (or cognitive strain) that is one of the warning signs that mobilizes System 2. However, when cognitive ease is present, the endorsements of System 1 are more readily accepted by System 2.
The feeling of cognitive ease is embedded in its own associative network of ideas internalized by System 1, and described on P.60.
When an idea/concept/communication has been repeated in the past, is displayed/communicated clearly, has been primed in advance, and conveyed when the subject is in a good mood, these all enhance the chances for the experience of cognitive ease. Likewise when the subject experiences cognitive ease associated with a stimuli, the subject is more likely to experience affects like the feeling of truth, feeling of familiarities, and feeling good/effortless.
This diagram has a variety of behavioral implications. It suggests that when cognitive ease is achieved, then statements are more likely to be believed. Given that mood, familiarity, clarity, priming affect cognitive ease, then one implication is that if you want your message to be believed by a person, you should use repetition, simple and clear language, have positive/likeable demeanor to affect mood, and prime (think “pre-suasion”) for your message.
This also implies the opposite result – when pre-conditions for cognitive ease are not realized (and cognitive strain is experienced) then System 2 is more likely to scrutinize the plausibility of conclusions. Consider the questions designed by Shane Frederick’s “Cognitive Reflection Test”
A bat and a ball cost $1.10
The bat costs one dollar more than the ball
How much does the bat cost?
All roses are flowers
Some flowers fade quickly
Therefore some roses fade quickly
In each case, the intuitive answer (1 dollar and yes) is the wrong answer that you can check with half a minute of deliberative thought. In fact, More than 50% of Harvard, MIT and Princeton students give the wrong answer – at other universities this is as high as 80%.
However “Quote TFS”.
Does this mean we are subject to inferior thinking due to our generally lazy System 2 behavior which is only incited due to a signal of strain from System 1. To the contrary, the ability of System 1 to sift through ideas in search of patterns of association that create cognitive ease is one of the drivers of creativity. See the findings on good feeling and the Remote Association Test on pages 68-69.
This suggests two distinct clusters of cognitive experiences.
Deliberation without Attention (DWoA)
Dijksterhuis, A., Bos, M. W., Nordgren, L. F., & van Baaren, R. B. (2006). On making the right choice: The deliberation-without-attention effect. Science, 311, 1005–1007.
Define
DWA = Deliberation With Attention
DWoA = Deliberation Without Attention
DWoA occurs when ...
\(a\) one considers a decision question;
\(b\) one spends some time thinking about something else (distraction);
\(c\) one returns to the decision and makes a decision.
The potential hypothesis of interest is that:
Deliberation places demands on a low capacity resource.
(Working memory is limited capacity. Deliberation is a System II process.)Deliberation can distort the weighting placed on different attributes.
Hypothesis: DWA better than DWoA on simple decisions;
DWA worse than DWoA on complex decisions.Why should this be?
DWA places heavy demands on limited working memory.
Complex problem overflows capacity of working memory.DWA makes errors by over-simplifying a complex decision.
DWoA processes complex decisions by associative processes
that do not require working memory.
Stimuli: 4 car descriptions:
Car A: 75% positive attributes, 25% negative attributes
Cars B and C: 50% positive attributes, 50% negative attributes
Car D: 25% positive attributes, 75% negative attributes
Experimental Design: \[Simple vs Complex\] x \[DWA vs DWoA\]
Simple Condition:
All cars had 4 attributes, e.g., mileage, safety, looks, etc.
Complex Condition: All cars had 12 attributes.
DWA: Look at the car descriptions; spend 4 minutes thinking
about what is the best choice; make choice.
DWoA: Look at the car descriptions; spend 4 minutes working
on a anagrams; make choice.
Results:
fadsfa
dafad
asdfd
Ddddww
Zzzz
Zd
Study 1: DWA and DWoA perform about the same on simple choice (4 aspects),
but DWoA performs better than DWA on complex choice (12 aspects).
Study 2: DWA has strong preference for best car in simple choice, but almost
no preference in complex choice. DWoA has strong preference
for best car in both simple and complex choice.
Interpretation of Studies 1 and 2:
- DWoA outperforms DWA on complex choices.
There is some evidence that DWA outperforms DWoA on simple choices.
Hypothesis: DWA places excessive demands on working memory, especially
when the decision is complex. This hurts performance on complex
decisions.
Shopping Study: Shopping Style & Item Deliberation
Shoppers answer questionnaires about how much time they spend thinking about their shopping choices.
Low Deliberation Shoppers (Unconscious Shoppers):
High Deliberation Shopper (Conscious Shoppers):
Shopping Items: Shampoo, CD, shoes, plane ticket, camera, room.
Low Deliberation Shopping Items, e.g., shampoo.
- Shoppers report considering few attributes before making a choice.
High Deliberation Shopping Items, e.g., camera.
Shoppers report considering many attributes before making a choice.
Experimental Design
The results are
Key result:
Low deliberation shoppers are more satisfied than high deliberation shoppers when the shopping item demands more deliberation.
.